Despite on-going complaints from both tenant and landlord groups, little reliable information existed for gauging the accuracy of the PIOC until 1990. In that year, RGB staff gained access to a new source of data which permitted independent verification of the PIOC's accuracy: Income and Expense (I&E) statements, collected annually by the Department of Finance from owners of "income producing" properties. These I&E statements contain detailed information on income and costs in rent stabilized buildings, and are particularly useful because they comprise both cross-sectional data, reflecting the condition of various types of rent stabilized buildings in a given year, and longitudinal data, which reflect changes in the condition of buildings which have filed I&E forms in at least two successive years.
Over the last six years Finance has provided the RGB with summary data for a random sample of rent stabilized properties. Samples in the first two studies were limited to 500 buildings, because RPIE files were not automated. Three years ago, following the computerization of all I&E filings, the sample size was increased to over 10,000 properties.
This is the sixth year that RGB staff has been able to use cross-sectional data and the fourth year that longitudinal figures have been used to monitor current conditions as well as trends in New York's rent stabilized housing. Because it traces actual income levels and costs (as reported by building owners) for the same properties over a number of years, longitudinal data is particularly useful for analyzing the recent performance of the PIOC in measuring changes in operating costs within the rent stabilized housing market.
The data used in this report was primarily summarized from 1994 RPIE forms returned to the Department of Finance by building owners. Longitudinal data encompasses properties which filed RPIE forms in both 1993 and 1994. However, analysis of filing dates indicates that RPIE averages reflect conditions occurring around July of the previous calendar year in question, so that this year's longitudinal study measures changes in costs and income from July 1992 to July 1993.
This year 13,363 and 11,425 buildings were respectively analyzed for the cross-sectional and longitudinal I&E studies. Figures were produced by matching a list of 39,000 rent stabilized properties registered with the New York State Division of Housing and Community Renewal (DHCR) with a list of buildings which had filed a 1994 RPIE statement (or both a 1993 and 1994 statement in the longitudinal sample). Buildings on the RGB list were excluded from both samples for the following reasons:
Three major steps were also taken to weed out inaccurate building information which could have distorted the final results:
After compiling both samples, Finance categorized sample data into "cells" reflecting particular types of rent stabilized buildings throughout the five boroughs (such as post-1946 rent stabilized buildings in Queens with 20-99 units) as they have done in the past.
Traditionally, average rents culled from RPIE filings tend to be lower than data on mean contract rents found in the triennial New York City Housing and Vacancy Survey (HVS). This disparity mainly stems from the fact that the I&E data accounts for vacancy and collection losses, in addition to reflecting rents collected over a 12-month period (the HVS is usually conducted in the first three months of any given year).
Using data from the 1993 HVS, the mean contract rent for all rent regulated apartments in 1993 ($574) exceeded the average rent from the 1994 RPIE data by roughly 6.1%[1] The mean contract rent in older pre-war apartments ($535) stood about 11% higher than 1994 RPIE average, while the 1993 mean contract rent for units built after 1946 ($695) was virtually the same as the 1994 RPIE average for such dwellings.
Similar gaps between HVS and RPIE data were observed in last year's I&E study, particularly for the pre-war sector, where mean contract rents exceeded average rent collections by 10%. If even a portion of these observed "gaps" between HVS and RPIE data reflect vacancy and collection losses, then it seems that older rent stabilized buildings continue to face much greater hardships than modern properties in the actual collection of their annual income.
It is also interesting to note the relationship between rent levels registered with the New York State Division of Housing and Community Renewal (DHCR) and rent collections reported by landlords in the I&E study. The gap between legal rents and rents actually collected may reflect a number of factors, including preferential rents, rents in controlled units, collection losses, and vacancy losses. Between 1988-1991, staff estimated that collected rents dropped from about 90% to 85% of registered rents. This gap closed slightly in recent years, as collections increased to 86% of registered rents between 1991-1992 and continued to grow to 87% from 1992-1993.
The disparity between collections and registered rents varied widely among the boroughs in 1993. Properties in Manhattan collected only 85% of the registered average, while buildings in Queens collected 90%. Collection rates in both Brooklyn and the Bronx stood at 88%.
Use of a sample exceeding 500,000 units allows reliable statistics to be calculated for rent in most of the building types encountered throughout New York's boroughs. The chart [ Buildings in Manhattan had Above Average Income and Costs in 1993] shows average rent for each of the building types.
Many owners of rent stabilized apartment buildings augment their income by selling services to their tenants as well as by renting ground floor commercial space. 1994 RPIE filings show an average gross income of $601 per rent stabilized unit, including the sales of services (e.g. laundry, garages/parking), as well as rent from commercial units. Such proceeds constituted roughly 11% of the total income earned by building owners in 1993. Manhattan owners especially benefit from commercial income, with 17% of their income coming from commercial units and services. The respective figures for the other boroughs are 6% in Queens and 5% in both Brooklyn and the Bronx.
The average monthly operating and maintenance cost for all rent stabilized units was $409 in 1993. Costs were substantially higher for Post '46 units ($502) and much lower for the pre-war stock ($373). In the boroughs costs parallel rents - lowest in the Bronx ($334) and highest in Manhattan ($506). The chart [ Large and Modern Rent Stabilized Buildings had Higher Operating Costs in 1993] shows costs according to building size and age.
Over the past six years, the Department of Finance and RGB staff have extensively scrutinized RPIE expense data for accuracy. Assessments of early samples indicated that more than half (55%) of "miscellaneous" costs were actually administrative or maintenance costs, while another 15% were not valid business expenses. Finance explored these findings further in 1992 by conducting thorough audits on the income and costs of forty six rent stabilized properties.
The auditors ultimately found that owners overstated O&M costs in RPIE filings by about 8%. Costs tended to be less accurately recorded in small (11-19 units) and medium (20-99 units) sized buildings (overstated by 13% and 9% respectively). Expenses in large (100+ units) buildings appeared to be more accurate (overstated on average by only 2%), but remain somewhat inconclusive since several owners of large stabilized properties refused to cooperate with Finance's assessors.
Expense reductions were concentrated in three categories: maintenance, administration, and miscellaneous costs. Maintenance had to be lowered by an average of 11% for all buildings, while administration and miscellaneous costs were respectively trimmed by approximately one-quarter (25%) and one-third (37%). Adjustment of 1994 RPIE data by the results of the 1992 audits reduces average O&M costs for stabilized buildings from $408 to $376.
Audit-adjusted monthly O&M costs for buildings without commercial units were about $30 lower ($346) than the average for all buildings. In 1993, RGB staff found that taxes accounted for 40% of the difference between "all-residential" buildings and all-stabilized buildings, with one-quarter of the remaining variance attributed to maintenance and labor expenses. This year taxes accounted for less than half (47%) of the difference while labor, maintenance and administrative costs accounted for nearly two-fifths (39%) of the total variation. Taxes, administrative and labor costs were respectively 15%, 10% and 7% lower on average for buildings without commercial space than for all stabilized properties.
In the past, I&E statistics have been computed by dividing the aggregate rent, income or costs reported in RPIE forms by the total number of apartments in responding properties. The decile distributions available this year are fundamentally different. Buildings in this year's cross-sectional sample were ranked three times, according to their average per-unit rents, incomes and costs. Once ranked, buildings were divided into ten sections of roughly 1300 buildings each. According to the way in which the buildings were ranked, the first decile represents buildings with the lowest average unit rents, incomes or costs, while buildings in the tenth decile have the highest average unit rents, incomes or costs.
Analysis of decile distributions produced interesting findings when compared with the unit-based averages for rents, income and costs reported earlier. Several commentators in the past have worried that the presence of a relatively small number of high rent apartments might distort unit-based averages upward. According to the rent, income and cost decile distributions computed this year, this seems to have happened across the board in 1993. Monthly rent in over 70% of stabilized buildings was less than the mean unit average of $541, while over 80% of the buildings in the cross sectional sample earned less monthly income per unit than the $601 mean unit average. Monthly expenses for units in more than 70% of buildings were less than the mean unit average cost of $409 reported earlier, as shown in the chart Monthly Costs in Most Stabilized Buildings Averaged Below $400 per Unit in 1993.
How did variation in these three variables affect the ratio of costs to income in stabilized buildings during 1993? Buildings in the first (lowest) rent decile spent an average of 125.7% of rent and 95.1% of income on operating costs, while buildings in the tenth (highest) rent decile spent an average of 69.6% of rent and 59.8% of income on expenses. Properties in the fifth rent decile used an average of 78.2% of rent and 71.8% of income for operating costs. More than 60% of these buildings spent greater proportions of their rent and income for expenses than the marketwide O&M-to-rent ratio of 75.6% and O&M-to-income ratio 68.1% (for unaudited costs).
Not surprisingly, buildings in the lower rent deciles tended to be smaller than those in the higher deciles. Similar trends were observed with the income and expense deciles. Buildings in the lowest income deciles spent greater proportions of both rent and income on operating costs than those in higher income deciles, while properties in the lowest expense deciles spent lower percentages of income and rent for costs than ones in higher expense deciles. While larger buildings dominated the higher income deciles, they also dominated the upper expense deciles, proving that building size alone is not a reliable predictor of profitability.
Last year, the Rent Guidelines Board heard considerable testimony concerning the need for a "minimum rent" ($400 per month was often mentioned), particularly for older buildings facing relatively high operating costs. Study of this year's cost deciles indicate that more than 70% of the stabilized buildings filing 1994 RPIE forms (roughly 9500 properties) had average unit operating costs below $400 per month, as the chart Monthly Costs in Most Stabilized Buildings Averaged Below $400 per Unit in 1993 illustrates. Moreover, only buildings in the eighth rent decile and above had average monthly operating costs equal or greater than $400 per unit.
Building size also affects the distribution of costs. Taxes, utilities, fuel and maintenance costs again dominated overall expenses in buildings of various sizes in 1993. Labor costs continued to be particularly associated with size, comprising much larger shares of total O&M costs in larger buildings. This may be due to the concentration of large modern (post-46) stabilized buildings in Manhattan, which tend to employ doormen. In contrast fuel and insurance decreased with size in 1993, probably due to efficiencies of scale realized by larger properties, particularly those with more than 100 units.
In buildings where expenses exceed income, unprofitability is both a function of abnormally high expenses (108% of the all-building average in 1993) and abnormally low rents (only 64% of the all-building average) and income (63% of the all building average in 1993), as the chart [ "Distressed" Buildings Suffered from Both Low Incomes and High Costs in 1993] shows. Most of the variance in unadjusted costs between these and other stabilized buildings was found in the insurance, fuel, maintenance and "miscellaneous" categories, which in these "distressed" buildings were respectively 115%, 129%, 130% and 195% of the stabilized average. Not surprisingly, these buildings also paid less property taxes (75% of the all-building average) than other stabilized structures in 1993. Taxes paid by "distressed" buildings in last year's cross sectional sample averaged 85% of the sector-wide mean. Unfortunately, available data cannot illuminate the cause of this gap.
Average rent increased by 3.8% from 1992 to 1993, slightly higher than the rise observed between 1991 and 1992 (3.5%). Rents in the post-46 sector went up 4% while charges in pre-47 buildings rose 3.8%. Large (100+ unit) buildings witnessed the fastest rent growth (3.9%) while rents in small (11-19 unit) buildings rose least (3.6%). Mid-sized (20-99 unit) properties experienced rent growth of 3.8% during 1993. In terms of both age and size, rents in small, post-war buildings actually decreased (by 4.8%) while those in small pre-war properties grew the most (4.8%).
1993 appears to be the first year that rent increases in Manhattan exceeded the marketwide average since New York City's economy slid into recession earlier in the decade (See chart Rents in Modern Manhattan Buildings Rebounded During 1993). In the recent past, above average rents and the City's weak economy prevented many owners of post-war properties in Manhattan from collecting all of the rent increases authorized by the RGB (the actual vacancy rate for modern buildings in the borough remained stable between 1991 and 1993). The "drought" that affected such buildings in 1991, during which rent collections actually declined, seems to have reversed somewhat during 1992 when rents in Manhattan's post-war stock grew by 1.5%, still far below the marketwide average. This recovery appears to have accelerated strongly in 1993, signalling renewed health in the higher end of New York's stabilized housing stock.
During the 1980's, rent collections accelerated faster than the RGB's expectations. This occurred again in 1993, as rent growth of 3.8% exceeded both the RGB's rent index (3.3%) and the increase observed in DHCR registered rents (2.9%) between 1992 and 1993.
Gross income (i.e. apartment rent, sales of services, and commercial rent) collected by owners between 1992 and 1993 increased by 3.4%, slightly less than growth in apartment rents. In contrast to previous years, income in post-46 units rose faster (3.6%) than in the pre-47 stock (3.4%). In terms of size, income grew fastest in large buildings (3.7%) and slowest in small ones (3.0%).
Among the various costs faced by building owners labor, administrative and utilities costs grew fastest (by respectively 5.1%, 3.1% and 2.3%) between 1992 and 1993. In contrast, insurance costs remained stable with only 0.4% growth, while fuel costs declined slightly (-0.4%). Growth in both taxes and maintenance expenses slackened to 1.6% each over the year, after surging by 7.7% and 4.2% respectively in 1992. Whether these trends reflect stagnating investment in rent stabilized properties, shifts away from undertaking in favor of routine maintenance, or merely inflation cannot be determined with certainty.
How do the changes in the I&E figures compare with the cost increases measured by the PIOC? Over the past few years, growth in PIOC-measured costs has consistently outpaced expense increases reported by building owners in RPIE data, as shown in the chart Cost Growth in the PIOC Outpaced the I&E Increase during 1993. During 1990, costs in the PIOC increased 9.6% while those reported to Finance grew by 7.1%. The following year, the PIOC rose 5.5% as RPIE costs went up 3.4%. This persistent gap closed in 1992, with costs in both the PIOC and RPIE filings growing by 4.2%. In 1993, the gap opened again, as costs in RPIE filings grew by 2.1% as opposed to 4% growth in expenses measured by the PIOC.
Comparison of I&E and PIOC data involves some distortion. Differences in the measurement of O&M components introduce error into comparisons between the two data sources. Additionally, many of the components examined in the PIOC are measured on an April-to-April basis, while most expense statements (88%) filed by landlords are based on the calendar year. Reconciling this difference requires use of a weighted average of two PIOC years to render figures resembling I&E data.
Analysis of PIOC and I&E data is further muddied by the fact that the two indices measure different things. Income and expense statements reflect actual expenditures incurred by landlords, while the PIOC heavily relies upon proxies to estimate actual shifts in O&M costs. Furthermore, the PIOC monitors the costs associated with maintaining properties to a constant standard of quality, while RPIE filings may reflect the investment or disinvestment patterns of building owners. Thus, rising O&M costs reported by RPIE filings may reflect price inflation, in which landlords are forced to spend more to maintain a given level of housing quality, or shifts in investment, where building owners change the quality of their buildings by spending more or less money to maintain them.
Despite those drawbacks, it is useful to make this comparison as one way of evaluating how well the PIOC methodology predicts changes in costs.
The chart PIOC Increases exceed I&E Data in Every Category Except Labor Costs illustrates the different growth rates reported by RPIE filings (for units in modern Manhattan buildings as well as all stabilized properties) and the PIOC for various costs between 1989 and 1993. Inclusion of modern Manhattan buildings allows quality to be rudimentally controlled, since such buildings have probably been the least likely to allow conditions to deteriorate due to competition for tenants. Between 1989 and 1993 the price index indicated a 25% increase in total O&M costs, while actual expenditures reported to Finance rose by 18% for all stabilized buildings and 16% for modern ones in Manhattan.
Reducing overall O&M expenses into component costs reveals exactly where the PIOC and RPIE data diverge. As the chart shows, the PIOC was most inaccurate in tracking both cyclical costs (fuel) and fairly stable ones (insurance, administration). In each of these cost categories, PIOC estimates were even more inaccurate for modern buildings than for the stabilized sector as a whole. Generally, the PIOC was most accurate in gauging increases for the most stable type of costs: taxes, utilities and labor. However, the gap between the price index and I&E cost data was larger for post-46 buildings than for all stabilized buildings. Overall, the PIOC for modern buildings registered 25% growth in costs between 1989 and 1993, as compared to 16% growth from I&E filings for such properties over the same period
In the case of insurance, the difference between PIOC and RPIE figures may reflect a decrease in the level of insurance used by building owners, although the PIOC does attempt to compensate for changes in coverage. The discrepancy in fuel costs may stem from the "degree-day" formula used to compute PIOC fuel costs, which may overemphasize changes in the weather. In addition, the I&E data does not account for lags between the consumption of fuel by building owners and the time they are billed by fuel providers.
Accurate assessment of the effectiveness of the PIOC and its various components is not possible in four years. However, data from modern Manhattan buildings shows that declining quality standards are probably not the cause of the disparity between the PIOC and I&E filings. Hopefully, future years will allow the RGB to trace the source of discrepancies between these two data sources and to maximize the performance of the PIOC in measuring operating cost changes.
Another price index commonly compared to the PIOC is the Consumer Price Index (CPI). The CPI, computed monthly by the Federal Bureau of Labor Statistics (BLS), tracks the price of a "market basket" of common goods thought to be purchased by the "typical" consumer. In existence since 1919, the CPI has recently been assailed as outdated and inaccurate. Critics claim that the CPI currently overstates price inflation by failing to account for two recent trends: 1) higher quality standards, particularly for electronics and computer related goods and 2) changing consumer spending patterns, which are intermittently surveyed by the BLS every five years.[2] According to its detractors, these drawbacks have led the CPI to overstate the nation’s inflation rate by at least one-sixth, and perhaps by as much as one-half.[3]
Although the PIOC and the CPI are similar methods of tracking prices, the two differ widely in what they measure. The CPI analyzes the prices of consumer goods, which are more volatile than the prices of the goods and services used for maintaining apartment buildings, which are analyzed by the PIOC. This stems from the fact that a large portion of consumer spending is "discretionary", involving the purchase of goods which are not strictly necessary, and which is directed by highly variable personal preferences. Most building owners indulge in much less "discretionary" spending when purchasing goods and services for the upkeep of their properties.
Additionally, technological changes have not increased the quality of goods tracked by the PIOC as they have affected consumer goods. The PIOC thus measures the prices of goods and services which have more stable quality standards than consumer goods, and which are purchased by building owners guided by more predictable spending habits than the average household.
One characteristic shared by the PIOC and the CPI might account for the variation witnessed between cost growth reported by the PIOC and I&E filings. While the "market baskets" of goods and services tracked by the PIOC and CPI are different, both are only intermittently updated. The Bureau of Labor Statistics reorganizes the CPI basket every five years, while that charted by the PIOC hasn’t been changed since 1982.[4] Although building owners exhibit less volatile spending habits than average consumers, it is possible that many of them are now spending their income on costs that did not exist when the PIOC was revamped in 1982. Such costs might include testing and abatement of lead paint (particularly in buildings built before 1960), recycling expenses, compliance with new municipal pollution and environmental laws and the effects of computerization on the performance of administrative duties. In light of these issues, staff is considering updating the PIOC market basket within the next year.
The percentage of buildings with an O&M to income ratio in excess of 100% declined from 11% to 10% of the roughly 11,000 buildings that filed RPIE forms in 1993 and 1994. Though there are slightly fewer buildings operating with an income ratio over 100%, the basic characteristics of these buildings did not differ from year to year. As reported in the cross-sectional study, these buildings have low average rents and high operating expenses.
